Images & Video · Blog post · source Aug 30, 2025
Main Image vs Lifestyle Image Test: Nano Banana vs OpenAI
Pick the image model for each job: main images vs lifestyle shots
For main images, the OpenAI model enlarged the product to fill the frame, cleaned the background, fixed the lighting and retouched the label, while Nano Banana kept the product as it was, with dull lighting and flat colours.
Per source Self-reported by the source, not independently verified.
Copy-paste prompt
Using the attached product photo, make an Amazon main image: pure white background (RGB 255,255,255), product filling about 85% of the frame, even studio lighting, label sharp and legible. Don't change the product's shape, colours, label text or proportions. Square, 2000x2000.
The source didn't publish its exact prompt. We wrote this starter so you can try the same workflow. Replace anything in [brackets].
What they did
Kamaljit Singh, who posts about Amazon listings, ran the same product and the same prompt through OpenAI's GPT image model and Gemini's Nano Banana (2.5 Flash) to see which makes better main images, then repeated the head-to-head a week later for lifestyle images.
How it went
For main images, the OpenAI model enlarged the product to fill the frame, cleaned the background, fixed the lighting and retouched the label, while Nano Banana kept the product as it was, with dull lighting and flat colours. For lifestyle images the result flipped: Nano Banana's looked natural and won by what he calls a decent margin, while OpenAI's looked slightly off.
We only report what the source says. Results are self-reported unless noted.
Worth knowing
A side-by-side judged by eye: he talks about CTR but reports no click data. The split is still a useful rule of thumb: models that redraw the product can win the main image, and models that preserve it are safer for lifestyle shots. Check any main image against the real product before upload.